APRICOT 2011 - MPLS-based Metro Ethernet Networks v1.0
APRICOT 2011 - MPLS-based Metro Ethernet Networks v1.0
APRICOT 2011 - MPLS-based Metro Ethernet Networks v1.0
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
A Tutorial<br />
Paresh Khatri<br />
Jan, 2010
<strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Paresh Khatri<br />
Director, Advanced Consulting Engineering<br />
2 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Agenda<br />
Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
Traditional <strong>Metro</strong> <strong>Ethernet</strong> networks<br />
Delivering <strong>Ethernet</strong> over <strong>MPLS</strong><br />
Summary<br />
Questions<br />
3 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
1. Introduction<br />
4 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Introduction<br />
Paresh Khatri (paresh.khatri@alcatel-lucent.com)<br />
Director – IP Competence Centre, APAC Solutions & Marketing, Alcatel-<br />
Lucent<br />
Key focus areas:<br />
Large-scale IP/<strong>MPLS</strong> networks<br />
L2/L3 VPNs<br />
Carrier <strong>Ethernet</strong><br />
Next-generation mobile backhaul networks<br />
Acknowledgements:<br />
Some figures and text are provided courtesy of the <strong>Metro</strong> <strong>Ethernet</strong> Forum (MEF)<br />
5 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
2. Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
6 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Agenda<br />
2. Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
2.1 Why <strong>Metro</strong> <strong>Ethernet</strong> ?<br />
2.2 Attributes of Carrier <strong>Ethernet</strong><br />
2.3 Carrier <strong>Ethernet</strong> Services defined by the MEF<br />
7 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
2.1 Why <strong>Metro</strong> <strong>Ethernet</strong> ?<br />
8 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
What is <strong>Metro</strong> <strong>Ethernet</strong> ?<br />
“… generally defined as the network that bridges or connects<br />
geographically separated enterprise LANs while also connecting across the<br />
WAN or backbone networks that are generally owned by service providers.<br />
The <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong> provide connectivity services across <strong>Metro</strong><br />
geography utilising <strong>Ethernet</strong> as the core protocol and enabling broadband<br />
applications”<br />
from “<strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong> – A Technical Overview” from the <strong>Metro</strong> <strong>Ethernet</strong> Forum<br />
9 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
Why <strong>Metro</strong> <strong>Ethernet</strong> ?<br />
Benefits both providers and customers in numerous ways …<br />
Packet traffic has now overtaken all other traffic types<br />
Need for rapid provisioning<br />
Reduced CAPEX/OPEX<br />
Increased and flexible bandwidth options<br />
Well-known interfaces and technology<br />
10 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
2.2 Attributes of Carrier <strong>Ethernet</strong><br />
11 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
The 5 Attributes of Carrier <strong>Ethernet</strong><br />
Carrier<br />
<strong>Ethernet</strong><br />
• Carrier <strong>Ethernet</strong> is a ubiquitous, standardized,<br />
carrier-class SERVICE defined by five<br />
attributes that distinguish Carrier <strong>Ethernet</strong><br />
from familiar LAN <strong>based</strong> <strong>Ethernet</strong><br />
• It brings the compelling business<br />
benefit of the <strong>Ethernet</strong> cost model<br />
to achieve significant savings<br />
• Standardized Services<br />
Carrier<br />
<strong>Ethernet</strong><br />
Attributes<br />
• Scalability<br />
• Service Management<br />
• Reliability<br />
• Quality of Service<br />
12 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
2.3 Carrier <strong>Ethernet</strong> Services defined by the MEF<br />
13 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Introduction to <strong>Metro</strong> <strong>Ethernet</strong> Services<br />
What do we mean by <strong>Metro</strong> <strong>Ethernet</strong> services ?<br />
Use of <strong>Ethernet</strong> access tails<br />
Provision of <strong>Ethernet</strong>-<strong>based</strong> services across the MAN/WAN<br />
Point-to-point<br />
Point-to-multipoint<br />
Multipoint-to-multipoint<br />
However, the underlying infrastructure used to deliver <strong>Ethernet</strong> services<br />
does NOT have to be <strong>Ethernet</strong> !!!<br />
Referred to as Carrier <strong>Ethernet</strong> services by the <strong>Metro</strong> <strong>Ethernet</strong> Forum<br />
The terms “Carrier <strong>Ethernet</strong>” and “<strong>Metro</strong> <strong>Ethernet</strong>” are used interchangeably in<br />
this presentation, but in the strict sense of the term, “Carrier <strong>Ethernet</strong>” refers to<br />
the carrier-grade evolution of “<strong>Metro</strong> <strong>Ethernet</strong>”<br />
14 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF Carrier <strong>Ethernet</strong> Terminology<br />
The User Network Interface (UNI)<br />
The UNI is the physical interface or port that is the demarcation<br />
between the customer and the service provider/Cable<br />
Operator/Carrier/MSO<br />
The UNI is always provided by the Service Provider<br />
The UNI in a Carrier <strong>Ethernet</strong> Network is a standard physical<br />
<strong>Ethernet</strong> Interface at operating speeds 10Mbs, 100Mbps, 1Gbps or<br />
10Gbps<br />
Carrier <strong>Ethernet</strong><br />
Network<br />
CE<br />
UNI<br />
CE: Customer Equipment, UNI: User Network Interface. MEF certified Carrier <strong>Ethernet</strong> products<br />
15 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF Carrier <strong>Ethernet</strong> Terminology<br />
The User Network Interface (UNI):<br />
MEF has defined two types of UNIs:<br />
MEF UNI Type I (MEF 13)<br />
– A UNI compliant with MEF 13<br />
– Manually configurable<br />
– Specified for existing <strong>Ethernet</strong> devices<br />
– Provides bare minimum data-plane connectivity services with no control-plane or<br />
management-plane capabilities.<br />
MEF UNI Type II (MEF 20)<br />
– Automatically configurable via E-LMI (allowing UNI-C to retrieve EVC status and<br />
configuration information from UNI-N)<br />
– Manageable via OAM<br />
Carrier <strong>Ethernet</strong><br />
Network<br />
CE<br />
UNI<br />
UNI<br />
CE: Customer Equipment, UNI: User Network Interface. MEF certified Carrier <strong>Ethernet</strong> products<br />
16 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF Carrier <strong>Ethernet</strong> Terminology<br />
Customer Equipment (CE) attaches to the <strong>Metro</strong> <strong>Ethernet</strong> Network<br />
(MEN) at the UNI<br />
Using standard <strong>Ethernet</strong> frames.<br />
CE can be<br />
Router or bridge/switch - IEEE 802.1 bridge<br />
Customer<br />
Edge<br />
(CE)<br />
User Network<br />
Interface<br />
(UNI)<br />
User Network<br />
Interface<br />
(UNI)<br />
Customer<br />
Edge<br />
(CE)<br />
<strong>Metro</strong><br />
<strong>Ethernet</strong><br />
Network<br />
17 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF <strong>Ethernet</strong> Services Model<br />
<strong>Ethernet</strong> Services “Eth” Layer<br />
Service Provider 1<br />
Service Provider 2<br />
<strong>Metro</strong> <strong>Ethernet</strong> Network<br />
<strong>Metro</strong> <strong>Ethernet</strong> Network<br />
Subscriber Site<br />
ETH<br />
UNI-C<br />
ETH<br />
UNI-N<br />
ETH<br />
UNI-N<br />
ETH<br />
UNI-N<br />
ETH<br />
UNI-N<br />
Subscriber Site<br />
ETH<br />
UNI-C<br />
UNI: User Network Interface, UNI-C: UNI-customer side, UNI-N network side<br />
NNI: Network to Network Interface, E-NNI: External NNI; I-NNI Internal NNI<br />
18 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF Carrier <strong>Ethernet</strong> Terminology<br />
<strong>Ethernet</strong> Virtual Connection (EVC)<br />
An <strong>Ethernet</strong> Service Instantiation<br />
Most commonly (but not necessarily) identified via a VLAN-ID<br />
Like Frame Relay and ATM PVCs or SVCs<br />
Connects two or more subscriber sites (UNI’s)<br />
Can multiplex multiple EVCs on the same UNI<br />
An association of two or more UNIs<br />
Prevents data transfer between sites that are not part of the same EVC<br />
19 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
MEF Carrier <strong>Ethernet</strong> Terminology<br />
<strong>Ethernet</strong> Virtual Connection (EVC)<br />
Three types of EVC:<br />
Point-to-Point EVC<br />
MEN<br />
UNI<br />
MEN<br />
UNI<br />
Multipoint-to-Multipoint EVC<br />
Leaf<br />
Leaf<br />
MEN<br />
Root<br />
Leaf<br />
Rooted-Multipoint EVC<br />
20 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Basic Carrier <strong>Ethernet</strong> Services<br />
E-LINE<br />
CE<br />
UNI<br />
Point-to-Point EVC<br />
UNI<br />
CE<br />
Point to Point<br />
Service Type used to<br />
create<br />
•<strong>Ethernet</strong> Private Lines<br />
•Virtual Private Lines<br />
•<strong>Ethernet</strong> Internet Access<br />
E-LAN<br />
CE<br />
UNI<br />
Multipoint EVC<br />
UNI<br />
CE<br />
Multi-Point to Multi-Point<br />
Service Type used to create<br />
•Multipoint Layer 2 VPNs<br />
•Transparent LAN Service<br />
E-TREE<br />
CE<br />
UNI<br />
UNI<br />
Rooted Multipoint EVC<br />
CE<br />
Point to Multi-Point<br />
•Efficient use of Service<br />
Provider ports<br />
•Foundation for Multicast<br />
networks e.g. IPTV<br />
UNI<br />
CE<br />
21 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
EVCs and Services<br />
In a Carrier <strong>Ethernet</strong> network, data is transported across Point-to-Point,<br />
Multipoint-to-Multipoint and Point-to-Multipoint EVCs according to the<br />
attributes and definitions of the E-Line, E-LAN and E-Tree services<br />
respectively.<br />
Point-to-Point EVC<br />
UNI<br />
UNI<br />
Carrier <strong>Ethernet</strong><br />
Network<br />
22 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Services Using E-Line Service Type<br />
<strong>Ethernet</strong> Private Line (EPL)<br />
Replaces a TDM Private line<br />
Dedicated UNIs for Point-to-Point connections<br />
Single <strong>Ethernet</strong> Virtual Connection (EVC) per UNI<br />
Storage Service<br />
Provider<br />
CE<br />
UNI<br />
UNI<br />
Carrier <strong>Ethernet</strong><br />
Network<br />
UNI<br />
CE<br />
ISP<br />
POP<br />
Internet<br />
Point-to-Point EVC<br />
UNI<br />
CE<br />
23 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Services Using E-Line Service Type<br />
<strong>Ethernet</strong> Virtual Private Line (EVPL)<br />
Replaces Frame Relay or ATM services<br />
Supports Service Multiplexed UNI<br />
(i.e. multiple EVCs per UNI)<br />
Allows single physical connection (UNI) to customer premise equipment for<br />
multiple virtual connections<br />
This is a UNI that must be configurable to support Multiple EVCs per UNI<br />
Service<br />
Multiplexed<br />
<strong>Ethernet</strong><br />
UNI<br />
UNI<br />
Carrier <strong>Ethernet</strong> Network<br />
CE<br />
CE<br />
UNI<br />
Multipoint-to-Multipoint EVC<br />
UNI<br />
CE<br />
24 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Services Using E-LAN Service Type<br />
<strong>Ethernet</strong> Private LAN and <strong>Ethernet</strong> Virtual Private LAN Services<br />
Supports dedicated or service-multiplexed UNIs<br />
Supports transparent LAN services and multipoint VPNs<br />
CE<br />
Service<br />
Multiplexed<br />
<strong>Ethernet</strong><br />
UNI<br />
UNI<br />
UNI<br />
Carrier<br />
<strong>Ethernet</strong><br />
Network<br />
UNI<br />
UNI<br />
CE<br />
Point-to-Multipoint EVC<br />
CE<br />
25 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Services Using E-Tree Service Type<br />
<strong>Ethernet</strong> Private Tree (EP-Tree) and <strong>Ethernet</strong> Virtual Private Tree (EVP-<br />
Tree) Services<br />
Enables Point-to-Multipoint Services with less provisioning than typical hub<br />
and spoke configuration using E-Lines<br />
Provides traffic separation between users with traffic from one “leaf” being allowed<br />
to arrive at one of more “roots” but never being transmitted to other “leaves”<br />
CE<br />
Root<br />
UNI<br />
Carrier <strong>Ethernet</strong> Network<br />
Leaf<br />
Leaf<br />
Leaf<br />
UNI<br />
UNI<br />
CE<br />
CE<br />
Rooted-Multipoint EVC<br />
UNI<br />
<strong>Ethernet</strong> Private Tree example<br />
CE<br />
26 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 1<br />
Name any two of the five attributes of Carrier<br />
<strong>Ethernet</strong> as defined by the <strong>Metro</strong> <strong>Ethernet</strong><br />
Forum.<br />
27 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3. Traditional <strong>Metro</strong> <strong>Ethernet</strong> networks<br />
28 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Agenda<br />
3. Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
3.1 Service Identification<br />
3.2 Forwarding Mechanism<br />
3.3 Resiliency and Redundancy<br />
3.4 Recent Developments<br />
3.5 Summary<br />
29 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Traditional methods of <strong>Ethernet</strong> delivery:<br />
<strong>Ethernet</strong> switching/bridging networks (802.1d/802.1q)<br />
Services identified by VLAN IDs/physical ports<br />
VLAN IDs globally significant<br />
Resiliency provided using variants of the Spanning Tree Protocol<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Agg<br />
Core<br />
Core<br />
Agg<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Access<br />
Agg<br />
Core<br />
Core<br />
Agg<br />
Access<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
<strong>Ethernet</strong> Switches<br />
30 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3.1 Service Identification<br />
31 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Service Identification:<br />
<strong>Ethernet</strong> switching/bridging networks<br />
First generation was <strong>based</strong> on IEEE 802.1q switches<br />
One obvious limitation was the VLAN ID space – the 12-bit VLAN ID allows a<br />
maximum of 4094 VLANs (VLANs 0 and 4095 are reserved). This limited the total<br />
number of services in any one switching/bridging domain.<br />
The other problem was that of customer VLAN usage – customers could not carry<br />
tagged traffic transparently across the network<br />
Payload<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
VLAN ID<br />
(12 bits)<br />
CFI (1 (1 bit)<br />
PCP(3 bits)<br />
0x8100<br />
(16 bits)<br />
Tag<br />
Control<br />
Information<br />
(TCI)<br />
Tag<br />
Protocol<br />
Identifer (TPID)<br />
32 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Service Identification :<br />
Q-in-Q (aka VLAN stacking, aka 802.1ad) comes to the rescue !<br />
Q-in-Q technology, which has now been standardised by the IEEE as 802.1ad<br />
(Provider Bridging), allowed the addition of an additional tag to customer <strong>Ethernet</strong><br />
frames – the S-tag. The S-tag (Service Tag) was imposed by the Service Provider<br />
and therefore, it became possible to carry customer tags (C-tags) transparently<br />
through the network.<br />
Customer<br />
Device<br />
Provider<br />
Bridge<br />
Payload<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
Payload<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
S-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
VLAN ID<br />
(12 bits)<br />
DEI (1 (1 bit)<br />
PCP(3 bits)<br />
0x88a8<br />
(16 bits)<br />
Tag<br />
Control<br />
Information<br />
(TCI)<br />
Tag<br />
Protocol<br />
Identifer (TPID)<br />
33 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Service Identification:<br />
Some important observations about Q-in-Q:<br />
This is not a new encapsulation format; it simply results in the addition of a second<br />
tag to the customer <strong>Ethernet</strong> frame, allowing any customer VLAN tags to be<br />
preserved across the network<br />
There is no change to the customer destination or source MAC addresses<br />
The number of distinct service instances within each Provider Bridging domain is<br />
still limited by the S-VLAN ID space i.e. 4094 S-VLANs. The difference is that<br />
customer VLANs can now be preserved and carried transparently across the<br />
provider network.<br />
34 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3.2 Forwarding Mechanism<br />
35 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Forwarding Mechanism:<br />
Dynamic learning methods used to build forwarding databases<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Agg<br />
Core<br />
Core<br />
Agg<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
CPE<br />
Access<br />
Access<br />
Agg<br />
Core<br />
Core<br />
Agg<br />
Access<br />
Access<br />
CPE<br />
CPE<br />
CPE<br />
MAC Learning Points<br />
36 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Forwarding Mechanism:<br />
Dynamic learning methods used to<br />
build forwarding databases<br />
Forwarding Database – E2<br />
MAC Interface<br />
MAC-A i6<br />
MAC-B i7<br />
MAC-C i6<br />
CPE<br />
(MAC A)<br />
i1<br />
Provider<br />
Switch<br />
E1<br />
i2<br />
i6<br />
Provider<br />
Switch<br />
E2<br />
i7<br />
CPE<br />
(MAC B)<br />
Forwarding Database – E1<br />
MAC Interface<br />
MAC-A i1<br />
MAC-B i2<br />
MAC-C i2<br />
i3<br />
Provider<br />
Switch<br />
C<br />
i4<br />
i8<br />
Provider<br />
Switch<br />
E3<br />
i9<br />
i5<br />
Forwarding Database – C<br />
MAC Interface<br />
MAC-A i3<br />
MAC-B i5<br />
MAC-C i4<br />
Forwarding Database – E3<br />
MAC Interface<br />
MAC-A i8<br />
CPE<br />
(MAC C)<br />
MAC-B<br />
MAC-C<br />
i8<br />
i9<br />
37 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Forwarding Mechanism:<br />
Dynamic learning methods used to build forwarding databases<br />
Data-plane process – there are no control-plane processes for discovering endpoint<br />
information<br />
In the worst case, ALL switches have forwarding databases that include ALL<br />
MAC addresses. This is true even for switches in the core of the network<br />
(Switch C in preceding example).<br />
Switches have limited resources for storing MAC addresses. This poses severe<br />
scaling issues in all parts of the network. VLAN-stacking does not help with this<br />
problem.<br />
On topology changes, forwarding databases are flushed and addresses need to be<br />
re-learned. While these addresses are re-learned, traffic to unknown destinations<br />
is flooded through the network, resulting in wasted bandwidth.<br />
38 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3.3 Resiliency and Redundancy<br />
39 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Resiliency and Redundancy<br />
Redundancy is needed in any network offering Carrier-grade <strong>Ethernet</strong> BUT<br />
loops are bad !!<br />
The Spanning Tree Protocol (STP) is used to break loops in bridged <strong>Ethernet</strong><br />
networks<br />
There have been many generations of the STP over the years<br />
All of these variants work by removing redundant links so that there is one, and<br />
only one, active path from each switch to every other switch i.e. all loops are<br />
eliminated. In effect, a minimum cost tree is created by the election of a root<br />
bridge and the subsequent determination of shortest-path links to the root bridge<br />
from every other bridge<br />
Bridges transmit special frames called Bridge Protocol Data Units (BPDUs) to<br />
exchange information about bridge priority, path costs etc.<br />
High Availability is difficult to achieve in traditional <strong>Metro</strong> <strong>Ethernet</strong><br />
networks.<br />
40 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Building the Spanning Tree …<br />
Root Bridge<br />
Switch<br />
A<br />
10<br />
Switch<br />
B<br />
Switch<br />
A<br />
10<br />
10<br />
Switch<br />
C<br />
20<br />
Switch<br />
D<br />
Switch<br />
B<br />
Switch<br />
C<br />
Switch<br />
D<br />
Rudimentary Traffic-Engineering Capabilities<br />
41 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
First generation of STP (IEEE802.1d-1998):<br />
Had a number of significant shortcomings:<br />
Convergence times – the protocol is timer-<strong>based</strong> with times in the order of 10s of<br />
seconds. After network topology changes (failure or addition of links), it could<br />
take up to 50s for the network to re-converge<br />
The protocol was VLAN-unaware, which meant that in an IEEE 802.1q network, all<br />
VLANs had to share the same spanning tree. This meant that there were network<br />
links that would not be utilised at all since they were placed into a blocked state.<br />
– Many vendors implemented their own, proprietary extensions to the protocol to<br />
allow the use of a separate STP instance per VLAN, allowing better link utilisation<br />
within the network<br />
There were many conditions which resulted in the inadvertent formation of loops in<br />
the network. Given the flooding nature of bridged <strong>Ethernet</strong>, and the lack of a TTLlike<br />
field in <strong>Ethernet</strong> frames, looping frames could loop forever.<br />
– There are numerous well-publicised instances of network meltdowns in Enterprise<br />
and Service Provider networks<br />
– A lot of service providers have been permanently scarred by the catastrophic effects<br />
of STP loops !<br />
42 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Newer generations of STP (IEEE802.1d-2004 – Rapid STP aka 802.1w):<br />
Some major improvements:<br />
Dependence on timers is reduced. Negotiation protocols have been introduced to<br />
allow rapid transitioning of links to a forwarding state<br />
The Topology Change process has been re-designed to allow faster recovery from<br />
topology changes<br />
Optimisations for certain types of direct and indirect link failures<br />
Convergence times are now down to sub-second in certain special cases but a lot of<br />
failure cases still require seconds to converge !<br />
But…<br />
The protocol was still VLAN-unaware, which meant that the issue of under-utilised<br />
links was still present<br />
43 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Newer generations of STP (IEEE802.1q-2003 – Multiple STP aka 802.1s):<br />
Built on top of RSTP<br />
Added VLAN awareness:<br />
Introduces the capability for the existence of multiple STP instances within the<br />
same bridged network<br />
Allows the association of VLANs to STP instances, in order to provide a (relatively)<br />
small number of STP instances, instead of using an instance per VLAN.<br />
Different STP instances can have different topologies, which allows much better<br />
link utilisation<br />
BUT<br />
The stigma associated with past failures is hard to remove…<br />
The protocol is fairly complicated, compared to its much simpler predecessors<br />
44 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3.4 Recent Developments<br />
45 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Provider Backbone Bridging<br />
Takes IEEE 802.1ad to the next level<br />
MAC-in-MAC technology:<br />
Customer <strong>Ethernet</strong> frames are encapsulated in a provider <strong>Ethernet</strong> frame<br />
Alleviates the MAC explosion problem<br />
Core switches no longer need to learn customer MAC addresses<br />
Does not address the STP issue, however.<br />
46 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Provider Backbone Bridging (PBB)<br />
<strong>Ethernet</strong> Technology being standardized in IEEE 802.1ah Task Group<br />
Designed to interconnect Provider Bridge <strong>Networks</strong> (PBN - IEEE 802.1ad)<br />
Adds a Backbone Header to a Customer/QinQ <strong>Ethernet</strong> Frame<br />
Provider Addressing for Backbone Forwarding<br />
New extended tag for Service Virtualization<br />
Standardization ongoing<br />
BEB:<br />
Backbone Edge Bridge<br />
Forward frames <strong>based</strong><br />
on backbone MAC<br />
addresses<br />
PBN<br />
PBBN<br />
PBN<br />
PBB<br />
BEB<br />
PBB<br />
BEB<br />
PBBN is <strong>Ethernet</strong> <strong>based</strong>:<br />
Connectionless Forwarding <strong>based</strong> on MAC Learning & Forwarding,<br />
Loop Avoidance <strong>based</strong> on STP,<br />
VLAN ID for Broadcast Containment<br />
47 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
IEEE 802.1ah Model for PBB – I and B Components<br />
Payload<br />
Payload<br />
Payload<br />
QinQ<br />
frame<br />
CMAC=Y<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
S-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
Customer FIB<br />
X->A1<br />
PBN<br />
(QinQ)<br />
Extended Service Tag<br />
Backbone VLAN ID<br />
I1<br />
Backbone MACs<br />
I2<br />
B2<br />
PBB PE2<br />
B4<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
S-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
I-SID<br />
Ethertype<br />
B-VID<br />
Ethertype<br />
B-SA<br />
B-DA<br />
Backbone FIBs<br />
A1->Port<br />
B5<br />
PBBN<br />
I1<br />
B3<br />
B6<br />
PBB<br />
frame<br />
Broadcast Containment<br />
MAC-<strong>based</strong>,<br />
Connectionless<br />
Forwarding<br />
B1<br />
A1 I2<br />
PBB PE1<br />
Ethertype<br />
C-VID<br />
Ethertype<br />
S-VID<br />
Ethertype<br />
C-SA<br />
C-DA<br />
Identifies the service instance inside PE<br />
I1<br />
PBN<br />
(QinQ)<br />
QinQ<br />
frame<br />
Customer FIB<br />
X->Port<br />
CMAC=X<br />
48 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
802.1ah Provider Backbone Bridge Encapsulation<br />
I-PCP = Customer Priority<br />
Payload<br />
I-DEI = Drop Elegibility<br />
Bits<br />
UCA = Use Customer Addresses<br />
I-SID = Service Instance ID<br />
3 1 1 3<br />
24<br />
I-PCP IDEI UCA Res I-SID<br />
C-TAG TCI<br />
q Etype = 81-00<br />
S – TAG TCI<br />
ad Etype = 88-a8<br />
C – SA<br />
C-TAG<br />
S-TAG<br />
C – DA<br />
I – TAG TCI<br />
ah Etype = 88-e7<br />
I-TAG<br />
2+4<br />
DEI<br />
p bits<br />
VLAN-ID<br />
B – TAG TCI<br />
ad Etype = 88-a8<br />
B – SA<br />
B-TAG<br />
2+2<br />
6+6<br />
B – DA<br />
22 (w/o FCS)<br />
49 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
3.5 Summary<br />
50 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Traditional <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong><br />
Summary of Issues:<br />
High Availability is difficult to achieve in networks running the Spanning<br />
Tree Protocol<br />
Scalability – IEEE 802.1q/802.1ad networks run into scalability limitations in<br />
terms of the number of supported services<br />
Customer <strong>Ethernet</strong> frames are encapsulated in a provider <strong>Ethernet</strong> frame<br />
QoS – only very rudimentary traffic-engineering can be achieved in bridged<br />
<strong>Ethernet</strong> networks.<br />
A lot of deployed <strong>Ethernet</strong> switching platforms lack carrier-class capabilities<br />
required for the delivery of Carrier <strong>Ethernet</strong> services<br />
New extensions in IEEE 802.1ah address some limitations such as the<br />
number of service instances and MAC explosion problems<br />
51 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 2<br />
Which IEEE standard defines Provider Bridging<br />
(Q-in-Q) ?<br />
52 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 3<br />
What is the size of the I-SID field in IEEE<br />
802.1ah?<br />
53 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4. Delivering <strong>Ethernet</strong> over <strong>MPLS</strong><br />
54 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Agenda<br />
4. Delivering <strong>Ethernet</strong> over <strong>MPLS</strong><br />
4.1 Introduction to <strong>MPLS</strong><br />
4.2 The Pseudowire Reference Model<br />
4.3 <strong>Ethernet</strong> Virtual Private Wire Service<br />
4.4 <strong>Ethernet</strong> Virtual Private LAN Service<br />
4.5 Scaling VPLS<br />
4.6 VPLS Topologies<br />
4.7 Resiliency Mechanisms<br />
55 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.1 Introduction to <strong>MPLS</strong><br />
56 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Delivering <strong>Ethernet</strong> over <strong>MPLS</strong><br />
<strong>MPLS</strong> Attributes<br />
Convergence: From “<strong>MPLS</strong> over everything” to “Everything over <strong>MPLS</strong>” !<br />
One network, multiple services<br />
Excellent virtualisation capabilities<br />
Today’s <strong>MPLS</strong> network can transport IP, ATM, Frame Relay and even TDM !<br />
Scalability<br />
<strong>MPLS</strong> is used in some of the largest service provider networks in the world<br />
Advanced Traffic Engineering capabilities using RSVP-TE<br />
Rapid recovery <strong>based</strong> on <strong>MPLS</strong> Fast ReRoute (FRR)<br />
Rapid restoration around failures by local action at the Points of Local Repair (PLRs)<br />
Sub-50ms restoration on link/node failures is a key requirement for carriers who are used to<br />
such performance in their SONET/SDH networks<br />
Feature-richness<br />
<strong>MPLS</strong> has 10 years of development behind it and continues to evolve today<br />
Layer 3 VPNs have already proven themselves as the killer app for <strong>MPLS</strong> – there is no<br />
reason why this success cannot be emulated by Layer 2 VPNs<br />
57 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> is truly Multi-Protocol<br />
The “Multiprotocol” nature of <strong>MPLS</strong>:<br />
<strong>MPLS</strong> is multiprotocol in terms of both the layers above and below it !<br />
The ultimate technology for convergence<br />
<strong>Ethernet</strong><br />
Frame<br />
Relay<br />
ATM<br />
TDM<br />
IP<br />
Etc.<br />
<strong>MPLS</strong><br />
<strong>Ethernet</strong><br />
Frame<br />
Relay<br />
ATM<br />
PoS<br />
PPP<br />
Etc.<br />
Physical<br />
58 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Virtualisation<br />
The virtualisation capabilities of <strong>MPLS</strong>:<br />
One common network supports multiple, different overlaid services<br />
PE<br />
PE<br />
P<br />
P<br />
PE<br />
P<br />
P<br />
<strong>MPLS</strong><br />
PE<br />
PE<br />
59 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Virtualisation<br />
The virtualisation capabilities of <strong>MPLS</strong>:<br />
One common network supports multiple, different overlaid services<br />
PE<br />
PE<br />
VPWS<br />
L3VPN<br />
VPLS<br />
PE<br />
<strong>MPLS</strong><br />
PE<br />
PE<br />
60 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Scalability<br />
<strong>MPLS</strong> Scalability:<br />
Service state is kept only on the Provider Edge devices<br />
The Provider (P) devices simply contain reachability information to each other and<br />
all PEs in the network<br />
The Provider Edge (PE) devices contain customer and service-specific state<br />
PE<br />
PE<br />
No<br />
customer<br />
or service<br />
state in<br />
the core<br />
P<br />
P<br />
P<br />
P<br />
PE<br />
<strong>MPLS</strong><br />
PE<br />
PE<br />
61 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Traffic-Engineering capabilities<br />
The Problem: consider example below – all mission-critical traffic between<br />
nodes A and Z has to use the path A-D-E-F-Z, while all other traffic uses the<br />
path A-B-C-Z.<br />
Other traffic<br />
B<br />
C<br />
A<br />
Z<br />
D E F<br />
Mission-critical traffic<br />
62 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
The IGP-<strong>based</strong> solution<br />
Use link metrics to influence traffic path<br />
It’s all or nothing – Traffic cannot be routed selectively<br />
Other solutions<br />
Policy-<strong>based</strong> routing – will work but is cumbersone to manage and has to be<br />
carefully crafted to avoid routing loops<br />
10<br />
B<br />
30<br />
C<br />
10<br />
A<br />
Z<br />
10<br />
D E F<br />
10 10<br />
10<br />
Mission-critical traffic<br />
Other traffic<br />
63 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
The <strong>MPLS</strong> solution<br />
Use constrained path routing to build Label Switched Paths (LSPs)<br />
Constrain LSP1 to use only the “orange” physical links<br />
Constrain LSP2 to use only the “blue” physical links<br />
At the PEs, map the mission-critical traffic to LSP2 and…<br />
…all other traffic to LSP1<br />
Other traffic<br />
LSP 1<br />
B<br />
C<br />
A<br />
Z<br />
D E F<br />
Mission-critical<br />
traffic<br />
LSP 2<br />
64 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Recovery from failures – typical IGP<br />
Step 1 – Detection of the failure<br />
One or more routers detect that a failure (link or node) has occurred<br />
Step 2 – Propagation of failure notification<br />
The router(s) detecting the failure inform other routers in the domain about the<br />
failure<br />
Step 3 – Recomputation of Paths/Routes<br />
All routers which receive the failure notification now have to recalculate new<br />
routes/paths by running SPF algorithms etc<br />
Step 4 – Updating of the Forwarding Table<br />
Once new routes are computed, they are downloaded to the routers’ forwarding<br />
table, in order to allow them to be used<br />
All of this takes time…<br />
65 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Failure and Recovery Example – IGP-<strong>based</strong><br />
What happens immediately after the link between C and Z fails ?<br />
Step 1 - Assuming a loss of signal (or similar physical indication) nodes C and Z<br />
immediately detect that the link is down<br />
Node A does not know that the link is down yet and keeps sending traffic destined<br />
to node Z to Node C. Assuming that node C has not completed step 4 yet, this<br />
traffic is dropped.<br />
10<br />
B<br />
20<br />
A<br />
10<br />
Z<br />
10<br />
C<br />
Direction of traffic flow<br />
66 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Failure and Recovery Example (continued) – IGP-<strong>based</strong><br />
Node C (and node Z) will be the first to recalculate its routing table and update its<br />
forwarding table (step 4).<br />
In the meantime, Node A does not know that the link is down yet and keeps sending<br />
traffic destined to node Z to Node C. Given that node C has completed step 4, it<br />
now believes (quite correctly) that the best path to Z is via node A. BUT – node A<br />
still believes that the best path to node Z is via node C so it sends the traffic right<br />
back to node C. We have a transient loop (micro-loop) ….<br />
The loop resolves itself as soon as node A updates its forwarding table but in the<br />
meantime, valuable packets have been dropped<br />
10<br />
B<br />
20<br />
A<br />
10<br />
Z<br />
Direction of traffic flow<br />
10<br />
C<br />
67 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Failure and Recovery Example (continued)<br />
Node A and all other nodes eventually update their forwarding tables and<br />
all is well again.<br />
But the damage is already done. . .<br />
10<br />
B<br />
20<br />
Direction of traffic flow<br />
A<br />
10<br />
Z<br />
10<br />
C<br />
68 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Recovery from failures – how can <strong>MPLS</strong> help ?<br />
RSVP-TE Fast Re-Route (FRR) pre-computes detours around potential failure<br />
points such as next-hop nodes and links<br />
When link or node failures occur, the routers (Points of Local Repair)<br />
directly connected to the failed link rapidly (sub-50ms) switch all traffic<br />
onto the detour paths.<br />
The network eventually converges and the head-end router (source of the<br />
traffic) switches traffic onto the most optimal path. Until that is done,<br />
traffic flows over the potentially sub-optimal detour path BUT the packet<br />
loss is kept to a minimum<br />
69 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Failure and Recovery Example – with <strong>MPLS</strong> FRR<br />
Node C pre-computes and builds a detour around link C-Z<br />
10<br />
B<br />
20<br />
A<br />
Bypass tunnel<br />
10<br />
Z<br />
10<br />
C<br />
Direction of traffic flow<br />
70 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>MPLS</strong> Traffic-Engineering<br />
Failure and Recovery Example – with <strong>MPLS</strong> FRR<br />
When link C-Z fails, node C reroutes traffic onto the detour tunnel<br />
Traffic does a U-turn but still makes it to the destination<br />
10<br />
B<br />
20<br />
A<br />
Direction of traffic flow<br />
10<br />
Z<br />
10<br />
C<br />
71 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 4<br />
What is the size of the <strong>MPLS</strong> label stack entry ?<br />
And the <strong>MPLS</strong> label itself ?<br />
72 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.2 The Pseudowire Reference Model<br />
73 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
The Pseudowire Reference Model<br />
Pseudowires:<br />
Key enabling technology for delivering <strong>Ethernet</strong> services over <strong>MPLS</strong><br />
Specified by the pwe3 working group of the IETF<br />
Originally designed for <strong>Ethernet</strong> over <strong>MPLS</strong> (Eo<strong>MPLS</strong>) – initially called Martini<br />
tunnels<br />
Now extended to many other services – ATM, FR, <strong>Ethernet</strong>, TDM<br />
Encapsulates and transports service-specific PDUs/Frames across a Packet<br />
Switched Network (PSN) tunnel<br />
The use of pseudowires for the emulation of point-to-point services is<br />
referred to as Virtual Private Wire Service (VPWS)<br />
IETF definition (RFC3985):<br />
“...a mechanism that emulates the essential attributes of a<br />
telecommunications service (such as a T1 leased line or Frame Relay)<br />
over a PSN. PWE3 is intended to provide only the minimum necessary<br />
functionality to emulate the wire with the required degree of<br />
faithfulness for the given service definition.”<br />
74 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
PWE3 Reference Model<br />
Generic PWE3 Architectural Reference Model:<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PSN Tunnel<br />
Pseudowire<br />
Emulated Service<br />
•Payload<br />
•Payload<br />
•Payload<br />
•PW Demultiplexer<br />
•PSN<br />
•Data Link<br />
•Physical<br />
75 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
PWE3 Terminology<br />
Pseudowire Terminology<br />
Attachment circuit (AC)<br />
The physical or virtual circuit attaching a CE to a PE.<br />
Customer Edge (CE)<br />
A device where one end of a service originates and/or terminates.<br />
Forwarder (FWRD)<br />
A PE subsystem that selects the PW to use in order to transmit a payload received on an AC.<br />
Packet Switched Network (PSN)<br />
Within the context of PWE3, this is a network using IP or <strong>MPLS</strong> as the mechanism for packet<br />
forwarding.<br />
Provider Edge (PE)<br />
A device that provides PWE3 to a CE.<br />
Pseudo Wire (PW)<br />
A mechanism that carries the essential elements of an emulated service from one PE to one or<br />
more other PEs over a PSN.<br />
PSN Tunnel<br />
A tunnel across a PSN, inside which one or more PWs can be carried.<br />
PW Demultiplexer<br />
Data-plane method of identifying a PW terminating at a PE.<br />
76 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Pseudowire Protocol Layering<br />
Pseudowire – Protocol Layering:<br />
The PW demultiplexing layer provides the ability to deliver multiple PWs<br />
over a single PSN tunnel<br />
•Payload<br />
<strong>Ethernet</strong> Frame<br />
•PW Label<br />
•PSN Label<br />
•Data Link<br />
•Physical<br />
<strong>Ethernet</strong> over <strong>MPLS</strong> PSN<br />
77 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.3 <strong>Ethernet</strong> Virtual Private Wire Service (VPWS)<br />
78 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires:<br />
Encapsulation specified in RFC4448 – “Encapsulation Methods for Transport<br />
of <strong>Ethernet</strong> over <strong>MPLS</strong> <strong>Networks</strong>”<br />
<strong>Ethernet</strong> pseudowires carry <strong>Ethernet</strong>/802.3 Protocol Data Units (PDUs) over<br />
an <strong>MPLS</strong> network<br />
Enables service providers to offer “emulated” <strong>Ethernet</strong> services over<br />
existing <strong>MPLS</strong> networks<br />
RFC4448 defines a point-to-point <strong>Ethernet</strong> pseudowire service<br />
Operates in one of two modes:<br />
Tagged mode - In tagged mode, each frame MUST contain at least one 802.1Q<br />
VLAN tag, and the tag value is meaningful to the two PW termination points.<br />
Raw mode - On a raw mode PW, a frame MAY contain an 802.1Q VLAN tag, but if it<br />
does, the tag is not meaningful to the PW termination points, and passes<br />
transparently through them.<br />
79 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires (continued):<br />
Two types of services:<br />
“port-to-port” – all traffic ingressing each attachment circuit is transparently<br />
conveyed to the other attachment circuit, where each attachment circuit is an<br />
entire <strong>Ethernet</strong> port<br />
“<strong>Ethernet</strong> VLAN to VLAN” – all traffic ingressing each attachment circuit is<br />
transparently conveyed to the other attachment circuit, where each attachment<br />
circuit is a VLAN on an <strong>Ethernet</strong> port<br />
– In this service instance, the VLAN tag may be stripped on ingress and<br />
then re-imposed on egress.<br />
– Alternatively, the VLAN tag may be stripped on ingress and a completely<br />
different VLAN ID imposed on egress, allowing VLAN re-write<br />
– The VLAN ID is locally significant to the <strong>Ethernet</strong> port<br />
80 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
PWE3 Reference Model for <strong>Ethernet</strong> VPWS<br />
PWE3 Architectural Reference Model for <strong>Ethernet</strong> Pseudowires<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PSN Tunnel<br />
Pseudowire<br />
Emulated Service<br />
•Payload<br />
•Payload<br />
•Payload<br />
•PW Demultiplexer<br />
•PSN<br />
•Data Link<br />
•Physical<br />
81 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> PWE3 Protocol Stack Reference Model:<br />
•Emulated<br />
•<strong>Ethernet</strong><br />
Emulated Service<br />
•Emulated<br />
•<strong>Ethernet</strong><br />
•PW Demultiplexer<br />
•PSN <strong>MPLS</strong><br />
•Data Link<br />
•Physical<br />
Pseudowire<br />
PSN Tunnel<br />
•PW Demultiplexer<br />
•PSN <strong>MPLS</strong><br />
•Data Link<br />
•Physical<br />
82 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> VPWS Example 1<br />
Example 1: <strong>Ethernet</strong> VPWS port-to-port (traffic flow from CE1 to CE2)<br />
PE1 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(port-to-port)<br />
PSN Label for PE2: 1029<br />
PW Label from PE2: 6775<br />
Port: 1/2/1<br />
Traffic Flow<br />
Port 1/2/1 PSN<br />
Port 3/2/0<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PE2 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(port-to-port)<br />
PSN Label for PE1: 4567<br />
PW Label from PE1: 10978<br />
Port: 3/2/0<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
•6775<br />
•1029<br />
•Data Link<br />
•Physical<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
83 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> VPWS Example 1<br />
Example 1: <strong>Ethernet</strong> VPWS port-to-port (traffic flow from CE2 to CE1)<br />
PE1 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(port-to-port)<br />
PSN Label for PE2: 1029<br />
PW Label from PE2: 6775<br />
Port: 1/2/1<br />
Traffic Flow<br />
Port 1/2/1 PSN<br />
Port 3/2/0<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PE2 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(port-to-port)<br />
PSN Label for PE1: 4567<br />
PW Label from PE1: 10978<br />
Port: 3/2/0<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
•10978<br />
•4567<br />
•Data Link<br />
•Physical<br />
•Payload<br />
VLAN tag<br />
SA SA<br />
DA<br />
84 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> VPWS Example 2<br />
Example 2: <strong>Ethernet</strong> VPWS VLAN-<strong>based</strong> (traffic flow from CE1 to CE2)<br />
PE1 Config:<br />
Service ID: 2000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(VLAN-100)<br />
PSN Label for PE2: 1029<br />
PW Label from PE2: 5879<br />
Port: 1/2/1 VLAN 100<br />
Traffic Flow<br />
Port 1/2/1 PSN<br />
Port 3/2/0<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PE2 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(VLAN-200)<br />
PSN Label for PE1: 4567<br />
PW Label from PE1: 21378<br />
Port: 3/2/0 VLAN 200<br />
•Payload<br />
VLAN tag --100<br />
SA SA<br />
DA<br />
•Payload<br />
SA SA<br />
DA<br />
•5879<br />
•1029<br />
•Data Link<br />
•Physical<br />
•Payload<br />
VLAN tag --200<br />
SA SA<br />
DA<br />
85 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> VPWS Example 2<br />
Example 2: <strong>Ethernet</strong> VPWS VLAN-<strong>based</strong> (traffic flow from CE2 to CE1)<br />
PE1 Config:<br />
Service ID: 2000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(VLAN-100)<br />
PSN Label for PE2: 1029<br />
PW Label from PE2: 5879<br />
Port: 1/2/1 VLAN 100<br />
Traffic Flow<br />
Port 1/2/1 PSN<br />
Port 3/2/0<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
PE2 Config:<br />
Service ID: 1000<br />
Service Type: <strong>Ethernet</strong> VPWS<br />
(VLAN-200)<br />
PSN Label for PE1: 4567<br />
PW Label from PE1: 21378<br />
Port: 3/2/0 VLAN 200<br />
•Payload<br />
VLAN tag --100<br />
SA SA<br />
DA<br />
•Payload<br />
SA SA<br />
DA<br />
•21378<br />
•4567<br />
•Data Link<br />
•Physical<br />
•Payload<br />
VLAN tag --200<br />
SA SA<br />
DA<br />
86 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
Signalling specified in RFC4447 – “Pseudowire Setup and Maintenance Using<br />
the Label Distribution Protocol (LDP)”<br />
The <strong>MPLS</strong> Label Distribution Protocol, LDP [RFC5036], is used for setting up<br />
and maintaining the pseudowires<br />
PW label bindings are distributed using the LDP downstream unsolicited mode<br />
PEs establish an LDP session using the LDP Extended Discovery mechanism a.k.a<br />
Targeted LDP or tLDP<br />
The PSN tunnels are established and maintained separately by using any of<br />
the following:<br />
The Label Distribution Protocol (LDP)<br />
The Resource Reservation Protocol with Traffic Engineering (RSVP-TE)<br />
Static labels<br />
87 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
LDP distributes FEC to label mappings using the PWid FEC Element (popularly<br />
known as FEC Type 128)<br />
Both pseudowire endpoints have to be provisioned with the same 32-bit identifier<br />
for the pseudowire to allow them to obtain a common understanding of which<br />
service a given pseudowire belongs to.<br />
0 1 2 3<br />
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| PWid (0x80) |C| PW type |PW info Length |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| Group ID |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| PW ID |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| Interface Parameter Sub-TLV |<br />
| " |<br />
| " |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
88 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
A new TLV, the Generalized PWid FEC Element (popularly known as FEC Type 129)<br />
has also been developed but is not widely deployed as yet<br />
The Generalized PWid FEC element requires that the PW endpoints be uniquely<br />
identified; the PW itself is identified as a pair of endpoints. In addition, the<br />
endpoint identifiers are structured to support applications where the identity of<br />
the remote endpoints needs to be auto-discovered rather than statically<br />
configured.<br />
89 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private Wire Service<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
The Generalized PWid FEC Element (popularly known as FEC Type 129)<br />
0 1 2 3<br />
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
|Gen PWid (0x81)|C| PW Type |PW info Length |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AGI Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ AGI Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AII Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ SAII Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AII Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ TAII Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
90 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 5<br />
What protocol is used to exchange pseudowire<br />
labels between provider edge routers ?<br />
91 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.4 <strong>Ethernet</strong> Virtual Private LAN Service (VPLS)<br />
92 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private LAN Service<br />
<strong>Ethernet</strong> VPLS:<br />
Two variants<br />
RFC4762 - Virtual Private LAN Service (VPLS) Using Label Distribution Protocol<br />
(LDP) Signaling. We will concentrate on this variant in the rest of this tutorial<br />
RFC4761 - Virtual Private LAN Service (VPLS) Using BGP for Auto-Discovery and<br />
Signaling<br />
93 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> Virtual Private LAN Service<br />
Definition:<br />
A VPLS creates an emulated private LAN segment for a given set of users.<br />
It creates a Layer 2 broadcast domain that is fully capable of learning and<br />
forwarding on <strong>Ethernet</strong> MAC addresses and that is closed to a given set of<br />
users. Multiple VPLS services can be supported from a single Provider Edge<br />
(PE) node.<br />
The primary motivation behind VPLS is to provide connectivity between<br />
geographically dispersed customer sites across MANs and WANs, as if they<br />
were connected using a LAN.<br />
The main intended application for the end-user can be divided into the<br />
following two categories:<br />
Connectivity between customer routers: LAN routing application<br />
Connectivity between customer <strong>Ethernet</strong> switches: LAN switching application<br />
94 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Benefits<br />
Benefits for the customer:<br />
Simplicity<br />
Behaves like an “ethernet switch in the sky”<br />
No routing interaction with the provider<br />
Clear demarcation between subscriber and provider<br />
Layer 3 agnostic<br />
Scalable<br />
Provider configures site connectivity only<br />
Hierarchy reduces number of sites touched<br />
Multi-site connectivity<br />
On the fly connectivity via <strong>Ethernet</strong> bridging<br />
95 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Topological Model<br />
Topological Model for VPLS (customer view)<br />
PSN<br />
CE 1 CE 2<br />
<strong>Ethernet</strong> Switch<br />
CE 3<br />
96 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Topological Model<br />
Topological Model for VPLS (provider view)<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
Emulated LAN<br />
PE 3<br />
CE 3<br />
Attachment<br />
Circuit<br />
97 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Constructing VPLS Services<br />
PSN Tunnels and Pseudowire Constructs for VPLS:<br />
Attachment Circuit<br />
Attachment Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 VB<br />
VB<br />
CE 2<br />
VB<br />
PSN (LSP) tunnel<br />
PE 3<br />
VB<br />
Virtual Bridge<br />
Instance<br />
CE 3<br />
Attachment Circuit<br />
Pseudowire<br />
98 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS PE Functions<br />
Provider Edge Functions:<br />
PE interfaces participating in a VPLS instance are able to flood, forward,<br />
and filter <strong>Ethernet</strong> frames, like a standard <strong>Ethernet</strong> bridged port<br />
Many forms of Attachment Circuits are acceptable, as long as they carry<br />
<strong>Ethernet</strong> frames:<br />
Physical <strong>Ethernet</strong> ports<br />
Logical (tagged) <strong>Ethernet</strong> ports<br />
ATM PVCs carrying <strong>Ethernet</strong> frames<br />
<strong>Ethernet</strong> Pseudowire<br />
Frames sent to broadcast addresses and to unknown destination MAC<br />
addresses are flooded to all ports:<br />
Attachment Circuits<br />
Pseudowires to all other PE nodes participating in the VPLS service<br />
PEs have the capability to associate MAC addresses with Pseudowires<br />
99 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS PE Functions<br />
Provider Edge Functions (continued):<br />
Address learning:<br />
Unlike BGP VPNs [RFC4364], reachability information is not advertised and<br />
distributed via a control plane.<br />
Reachability is obtained by standard learning bridge functions in the data plane.<br />
When a packet arrives on a PW, if the source MAC address is unknown, it is<br />
associated with the PW, so that outbound packets to that MAC address can be<br />
delivered over the associated PW.<br />
When a packet arrives on an AC, if the source MAC address is unknown, it is<br />
associated with the AC, so that outbound packets to that MAC address can be<br />
delivered over the associated AC.<br />
100 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Signalling<br />
VPLS Mechanics:<br />
Bridging capable PE routers are<br />
connected with a full mesh of <strong>MPLS</strong><br />
LSP tunnels<br />
Per-Service pseudowire labels are<br />
negotiated using RFC 4447<br />
techniques<br />
Replicates unknown/broadcast<br />
traffic in a service domain<br />
MAC learning over tunnel & access<br />
ports<br />
Separate FIB per VPLS for private<br />
communication<br />
Attachment<br />
Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
Full mesh of<br />
LSP tunnels<br />
VPLS<br />
Service<br />
PE 3<br />
CE 3<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
101 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Signalling<br />
Tunnel establishment<br />
LDP:<br />
<strong>MPLS</strong> paths <strong>based</strong> on IGP reachability<br />
RSVP: traffic engineered <strong>MPLS</strong> paths<br />
with bandwidth & link constraints,<br />
and fast reroute alternatives<br />
Pseudowire establishment<br />
LDP: point-to-point exchange of PW<br />
ID, labels, MTU<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
PSN<br />
PE 1 PE 2<br />
CE 1 CE 2<br />
VPLS<br />
Service<br />
PE 3<br />
Full mesh of<br />
LSP tunnels<br />
CE 3<br />
Attachment<br />
Circuit<br />
102 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Signalling<br />
A full mesh of pseudowires is established between all PEs<br />
participating in the VPLS service:<br />
Each PE initiates a targeted LDP session to the far-end System IP (loopback)<br />
address<br />
Tells far-end what PW label to use when sending packets for each service<br />
Attachment<br />
Circuit<br />
Attachment<br />
Circuit<br />
PE 1 PE 2<br />
PSN<br />
CE 1 VB<br />
VB CE 2<br />
VB<br />
PSN (LSP) tunnel<br />
PE 3<br />
VB<br />
Virtual Bridge<br />
Instance<br />
Pseudowire<br />
CE 3<br />
Attachment<br />
Circuit<br />
103 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Signalling<br />
Why a full mesh of pseudowires?<br />
If the topology of the VPLS is not restricted to a full mesh, then it may<br />
be that for two PEs not directly connected via PWs, they would have to<br />
use an intermediary PE to relay packets<br />
A loop-breaking protocol, such as the Spanning Tree Protocol, would be<br />
required<br />
With a full-mesh of PWs, every PE is now directly connected to every<br />
other PE in the VPLS via a PW; there is no longer any need to relay<br />
packets<br />
The loop-breaking rule now becomes the "split horizon" rule, whereby a<br />
PE MUST NOT forward traffic received from one PW to another in the<br />
same VPLS mesh<br />
Does this remind you of a similar mechanism used in IP networks ? The ibgp<br />
full-mesh !<br />
104 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Pseudowire Signalling<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
Signalling specified in RFC4447 – “Pseudowire Setup and Maintenance Using<br />
the Label Distribution Protocol (LDP)”<br />
The <strong>MPLS</strong> Label Distribution Protocol, LDP [RFC5036], is used for setting up<br />
and maintaining the pseudowires<br />
PW label bindings are distributed using the LDP downstream unsolicited mode<br />
PEs establish an LDP session using the LDP Extended Discovery mechanism a.k.a<br />
Targeted LDP or tLDP<br />
The PSN tunnels are established and maintained separately by using any of<br />
the following:<br />
The Label Distribution Protocol (LDP)<br />
The Resource Reservation Protocol with Traffic Engineering (RSVP-TE)<br />
Static labels<br />
105 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Pseudowire Signalling<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
LDP distributes FEC to label mappings using the PWid FEC Element (popularly<br />
known as FEC Type 128)<br />
Both pseudowire endpoints have to be provisioned with the same 32-bit identifier<br />
for the pseudowire to allow them to obtain a common understanding of which<br />
service a given pseudowire belongs to.<br />
0 1 2 3<br />
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| PWid (0x80) |C| PW type |PW info Length |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| Group ID |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| PW ID |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| Interface Parameter Sub-TLV |<br />
| " |<br />
| " |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
106 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Pseudowire Signalling<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
A new TLV, the Generalized PWid FEC Element (popularly known as FEC Type 129)<br />
has also been developed but is not widely deployed as yet<br />
The Generalized PWid FEC element requires that the PW endpoints be uniquely<br />
identified; the PW itself is identified as a pair of endpoints. In addition, the<br />
endpoint identifiers are structured to support applications where the identity of<br />
the remote endpoints needs to be auto-discovered rather than statically<br />
configured.<br />
107 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Pseudowire Signalling<br />
<strong>Ethernet</strong> Pseudowires – Setup and Maintenance:<br />
The Generalized PWid FEC Element (popularly known as FEC Type 129)<br />
0 1 2 3<br />
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
|Gen PWid (0x81)|C| PW Type |PW info Length |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AGI Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ AGI Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AII Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ SAII Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
| AII Type | Length | Value |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
~ TAII Value (contd.) ~<br />
| |<br />
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+<br />
108 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
<strong>Ethernet</strong> VPLS Signalling Example<br />
PE1 Config:<br />
Service ID: 1001<br />
Service Type: <strong>Ethernet</strong> VPLS<br />
PSN Label for PE2: 1029<br />
PSN Label for PE3: 9178<br />
PW Label from PE2: 6775<br />
PW Label from PE3: 10127<br />
Port: 1/2/1<br />
M1<br />
Port<br />
1/2/1<br />
VB<br />
PSN<br />
VB<br />
Port<br />
3/2/0<br />
PE 1 PE 2<br />
PE2 Config:<br />
Service ID: 1001<br />
Service Type: <strong>Ethernet</strong> VPLS<br />
PSN Label for PE1: 4567<br />
PSN Label for PE3: 11786<br />
PW Label from PE1: 10978<br />
PW Label from PE3: 4757<br />
Port: 3/2/0<br />
M2<br />
PE3 Config:<br />
Service ID: 1001<br />
Service Type: <strong>Ethernet</strong> VPLS<br />
PSN Label for PE1: 6668<br />
PSN Label for PE2: 12812<br />
PW Label from PE1: 4568<br />
PW Label from PE3: 10128<br />
Port: 4/1/2<br />
PE 3<br />
VB<br />
M3<br />
Port 4/1/2<br />
109 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Packet Walkthrough and MAC Learning Example<br />
Packet Walkthrough for<br />
VPLS Service-id 1001<br />
M1<br />
Port<br />
1/2/1<br />
VB<br />
PSN<br />
VB<br />
Port<br />
3/2/0<br />
PE 1 PE 2<br />
M2<br />
Forwarding Database – PE 1<br />
Forwarding Database – PE 2<br />
MAC<br />
M2<br />
Location<br />
Remote<br />
Mapping<br />
PW to PE2<br />
VB<br />
MA<br />
C<br />
M2<br />
Locatio<br />
n<br />
Local<br />
Mapping<br />
Port 3/2/0<br />
PE 3<br />
Port 4/1/2<br />
Send a packet from M2 to M1<br />
M3<br />
- PE2 learns that M2 is reached on Port 3/2/0<br />
- PE2 floods to PE1 with PW-label 10978 and PE3 with PW-label 4757<br />
- PE1 learns from the PW-label 10978 that M2 is behind PE2<br />
- PE1 sends on Port 1/2/1<br />
- PE3 learns from the PW-label 4757 M2 is behind PE2<br />
- PE3 sends on Port 4/1/2<br />
- M1 receives packet<br />
Forwarding Database – PE 3<br />
MAC Location Mapping<br />
M2 Remote PW to PE2<br />
110 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
VPLS Packet Walkthrough and MAC Learning Example<br />
(cont.)<br />
Packet Walkthrough for<br />
VPLS Service-id 1001<br />
M1<br />
Port<br />
1/2/1<br />
VB<br />
PSN<br />
VB<br />
Port<br />
3/2/0<br />
PE 1 PE 2<br />
M2<br />
Forwarding Database – PE 1<br />
Forwarding Database – PE 2<br />
MAC<br />
M1<br />
M2<br />
Location<br />
Local<br />
Remote<br />
Mapping<br />
Port 1/2/1<br />
PW to PE2<br />
PE 3<br />
VB<br />
Port 4/1/2<br />
MA<br />
C<br />
M1<br />
M2<br />
Locatio<br />
n<br />
Remote<br />
Local<br />
Mapping<br />
PW to PE1<br />
Port 3/2/0<br />
Reply with a packet from M1 to M2<br />
- PE1 learns M1 is on Port 1/2/1<br />
- PE1 knows that M2 is reachable via PE2<br />
- PE1 sends to PE2 using PW-label 6775<br />
M3<br />
- PE2 knows that M2 is reachable on Port 3/2/0 and so it sends it out that port<br />
- M2 receives packet<br />
111 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Audience Question 6<br />
If a full-mesh VPLS is set up between 5 provider<br />
edge routers, how many pseudowires need to be<br />
configured ?<br />
112 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.5 Scaling VPLS<br />
113 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Hierarchical-VPLS (H-VPLS)<br />
Introduces hierarchy in the base VPLS solution to provide scaling &<br />
operational advantages<br />
Extends the reach of a VPLS using spokes, i.e., point-to-point<br />
pseudowires or logical ports<br />
M-1<br />
PE-1<br />
VB<br />
VPLS<br />
PE-2<br />
VB<br />
M-3<br />
VB<br />
M-5<br />
M-6<br />
MTU-1<br />
VB<br />
VB<br />
PE-3<br />
114 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Hierarchical VPLS<br />
How is a spoke useful?<br />
Scales signalling<br />
Full-mesh between MTUs is reduced to full-mesh between PEs and<br />
single PW between MTU and PE<br />
Scales replication<br />
Replication at MTU is not required<br />
Replication is reduced to what is necessary between PEs<br />
Simplifies edge devices<br />
Keeps cost down because PEs can be replaced with MTUs<br />
Enables scalable inter-domain VPLS<br />
Single spoke to interconnect domains<br />
115 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Scalability: Signalling<br />
Full-mesh between PEs is reduced to full-mesh between PEs and<br />
single spoke between MTU and PE<br />
Spoke PWs<br />
Mesh PWs<br />
Mesh PWs<br />
116 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Scalability: Replication<br />
Flat architecture replication is reduced to distributed replication<br />
117 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Scalability: Configuration<br />
Full mesh configuration<br />
is significantly reduced<br />
118 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topological Extensibility: <strong>Metro</strong> Interconnect<br />
<strong>Metro</strong><br />
IP / <strong>MPLS</strong><br />
Network<br />
ISP<br />
IP / <strong>MPLS</strong><br />
Core Network<br />
<strong>Metro</strong><br />
IP / <strong>MPLS</strong><br />
Network<br />
119 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topological Extensibility: Inter-AS Connectivity<br />
Provider hand-off can be<br />
q-tagged or q-in-q port<br />
Pseudowire spoke<br />
Provider A<br />
IP / <strong>MPLS</strong><br />
Network<br />
Provider B<br />
IP / <strong>MPLS</strong><br />
Network<br />
120 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.6 VPLS Topologies<br />
121 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topologies: Mesh<br />
PE-1<br />
PE-2<br />
PE-4<br />
PE-3<br />
122 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topologies: Hierarchical<br />
PE-1<br />
PE-2<br />
PE-4<br />
PE-3<br />
123 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topologies: Dual-homing<br />
PE-1<br />
PE-2<br />
PE-4<br />
PE-3<br />
124 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Topologies: Ring<br />
A full mesh would have too<br />
many duplicate packets<br />
Each PE has a spoke to the<br />
next PE in the VPLS<br />
Packets are flooded into the<br />
adjacent spokes and to all<br />
VPLS ports<br />
When MACs are learned,<br />
packets stop at the owning<br />
PE<br />
PE-1<br />
PE-2<br />
PE-3<br />
PE-4<br />
PE-6<br />
PE-5<br />
125 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.7 Resiliency Mechanisms<br />
126 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Agenda<br />
4.7. Resiliency Mechanisms<br />
4.7.1 Multi-Chassis LAG (MC-LAG)<br />
4.7.2 Redundancy with VPLS<br />
4.7.3 Pseudo-wire Redundancy with MC-LAG<br />
4.7.4 Multi-Segment Pseudo-wires<br />
127 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.7.1 Multi-Chassis LAG (MC-LAG)<br />
128 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG: What is it ?<br />
Standard LAG<br />
Traffic distributed via hash algorithm<br />
Maintains packet sequence per “flow”<br />
Based on packet content or SAP/service ID<br />
LAG 1 LAG 1<br />
What if one system fails…<br />
Link Aggregation Control Protocol (LACP)<br />
Introduce LAG redundancy to TWO systems<br />
IEEE Std 802.3-2002_part3 (formerly in 802.3ad)<br />
Multi-Chassis LAG (MC-LAG)<br />
system MAC and priority<br />
system MAC and priority<br />
administrative key<br />
administrative key<br />
Consistent port capabilities (e.g. speed, duplex)<br />
129 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG: How does it work ?<br />
Multi-chassis LAG<br />
Active<br />
lag 1 lacp-key 1<br />
lag 1 lacp-key 1<br />
system-id 00:00:00:00:00:01<br />
system-id 00:00:00:00:00:01<br />
system-priority 100<br />
system-priority 100<br />
Standard LAG<br />
MC-LAG<br />
Edge<br />
device<br />
LAG 1<br />
(subgroup)<br />
(subgroup)<br />
LAG 1<br />
Multi-chassis LAG<br />
control protocol<br />
Provider<br />
Network<br />
LAG<br />
1<br />
MC-LAG<br />
out of sync<br />
in LACPDUs<br />
Standby<br />
lag 1 lacp-key 1<br />
lag 1 lacp-key 1<br />
system-id 00:00:00:00:00:01<br />
system-id 00:00:00:00:00:01<br />
system-priority 100<br />
system-priority 100<br />
LACP<br />
MC-LAG on a SAP<br />
130 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG: How does it work ?<br />
Multi-chassis LAG failover<br />
Active<br />
Standard LAG<br />
MC-LAG<br />
msg<br />
Edge<br />
device<br />
LAG 1<br />
(subgroup)<br />
(subgroup)<br />
LAG 1<br />
Multi-chassis LAG<br />
control protocol<br />
Provider<br />
Network<br />
LAG<br />
1<br />
MC-LAG<br />
in out sync of sync<br />
in LACP LACPDUs message<br />
Standby Active<br />
LACP<br />
131 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.7.2 Redundancy with VPLS<br />
132 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Redundancy at the VPLS edge: MC-LAG<br />
Active<br />
Triggered by Phy/<br />
LACP/802.3ah<br />
failure detection<br />
MC-LAG<br />
MAC<br />
withdraw<br />
Standard<br />
LAG<br />
VPLS<br />
LAG<br />
MC-LAG<br />
Standby Active<br />
133 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Redundancy Applications for VPLS w/MC-LAG<br />
Network Edge<br />
L2/L3 CPE for business services L2 DSLAM/BRAS for triple-play services<br />
CE<br />
Active<br />
Standby<br />
MC-LAG<br />
MC-LAG<br />
Provider<br />
Network<br />
DSLAM<br />
Active<br />
MC-LAG<br />
Provider<br />
Network<br />
Standby<br />
MC-LAG<br />
Inter-metro Connectivity<br />
Single active path<br />
MC-LAG<br />
Active<br />
MC-LAG<br />
Selective MAC<br />
withdraw for<br />
faster convergence<br />
VPLS<br />
Full<br />
Mesh<br />
MC-LAG<br />
Full<br />
Mesh<br />
VPLS<br />
MC-LAG<br />
Standby<br />
MC-LAG<br />
134 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.7.3 Pseudo-wire Redundancy with Multi-chassis LAG<br />
135 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Pseudowire Redundancy<br />
VLL<br />
PW<br />
VLL<br />
• Tunnel redundancy<br />
Access<br />
Node<br />
Tunnel bypass<br />
Access<br />
Node<br />
VLL<br />
• PW redundancy<br />
• Single edge redundancy<br />
Access<br />
Node<br />
LAG<br />
Redundant PW<br />
Access<br />
Node<br />
VLL<br />
• PW redundancy<br />
• Dual edge redundancy LAG LAG<br />
Access<br />
Node<br />
Redundant PW<br />
Access<br />
Node<br />
136 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Combining MC-LAG with Pseudowire Redundancy<br />
Extends L2 point-to-point redundancy across the network<br />
MC-LAG status propagated<br />
to local PW end points<br />
Local PW status signaled via T-LDP<br />
PW showing both ends<br />
active preferred for forwarding<br />
Active<br />
Active<br />
Access<br />
Node<br />
MC-<br />
LAG<br />
Standby<br />
Redundant PW<br />
Active<br />
Access<br />
Node<br />
VLL service terminates on different devices<br />
137 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with Pseudo-Wire Redundancy:<br />
How does it work ?<br />
VLL<br />
PW<br />
VLL<br />
• PW redundancy<br />
• Single edge redundancy<br />
Access<br />
Node<br />
LAG<br />
Access<br />
Node<br />
138 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with PW Redundancy:<br />
How does it work ?<br />
LAG to PWs<br />
Traffic path<br />
B<br />
A<br />
SAP<br />
epipe<br />
Standard<br />
LAG<br />
Active<br />
Standby<br />
LAG<br />
MC-LAG<br />
MC-LAG<br />
A<br />
C<br />
epipe<br />
epipe<br />
B<br />
D<br />
PW<br />
PW<br />
Active Active<br />
Standby Active<br />
PWs<br />
PW<br />
PW<br />
X<br />
epipe<br />
Y<br />
SAP<br />
D<br />
C<br />
139 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with PW Redundancy:<br />
How does it work ?<br />
LAG to PWs : LAG link failure<br />
Traffic path<br />
B<br />
A<br />
SAP<br />
epipe<br />
Standard<br />
LAG<br />
Active<br />
Standby Active<br />
LAG<br />
MC-LAG<br />
MC-LAG<br />
A<br />
C<br />
epipe<br />
epipe<br />
B<br />
D<br />
S SDP<br />
S SDP<br />
Active Active<br />
Active Standby Active<br />
PWs<br />
S SDP<br />
S SDP<br />
X<br />
epipe<br />
Y<br />
SAP<br />
D<br />
New Traffic path<br />
C<br />
140 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with Pseudo-Wire Redundancy:<br />
How does it work ?<br />
VLL<br />
PW<br />
VLL<br />
• PW redundancy<br />
• Dual edge redundancy<br />
Access<br />
Node<br />
LAG<br />
LAG<br />
Access<br />
Node<br />
141 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with PW Redundancy:<br />
How does it work ?<br />
LAG to PWs to LAG<br />
B<br />
D<br />
MC-LAG<br />
PW<br />
Active<br />
Standby<br />
PW<br />
MC-LAG<br />
A<br />
Active<br />
PW<br />
Active<br />
Standby<br />
Pw<br />
Standby<br />
F<br />
Standard<br />
LAG<br />
Standby<br />
Active<br />
Standard<br />
LAG<br />
LAG<br />
Standby<br />
MC-LAG<br />
PW<br />
PW<br />
Standby<br />
Active<br />
PW<br />
PW<br />
MC-LAG<br />
Active<br />
LAG<br />
C<br />
PWs<br />
E<br />
Traffic path<br />
142 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-chassis LAG with PW Redundancy:<br />
How does it work ?<br />
LAG to PWs to LAG : Network device failure<br />
B<br />
D<br />
MC-LAG<br />
PW<br />
Active<br />
Standby<br />
PW<br />
MC-LAG<br />
A<br />
Active<br />
PW<br />
Active<br />
Standby Standby<br />
PW<br />
Standby<br />
F<br />
Standard<br />
LAG<br />
Standby Active<br />
Active<br />
Standard<br />
LAG<br />
LAG<br />
Active Standby<br />
MC-LAG<br />
PW<br />
PW<br />
Standby Active<br />
Active<br />
PW<br />
PW<br />
MC-LAG<br />
Active<br />
LAG<br />
C<br />
PWs<br />
E<br />
Traffic path<br />
New Traffic path<br />
143 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
4.7.4 Multi-segment Pseudo-wires<br />
144 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-segment Pseudo-wire – Motivation<br />
<strong>Ethernet</strong> VLL with SS-PW<br />
<strong>MPLS</strong> tunnel<br />
SS-PW<br />
PE<br />
T-LDP<br />
PE<br />
CE<br />
CE<br />
<strong>MPLS</strong><br />
P<br />
P<br />
<strong>MPLS</strong><br />
CE<br />
PE<br />
T-LDP<br />
<strong>MPLS</strong><br />
PE<br />
CE<br />
T-LDP<br />
<strong>MPLS</strong><br />
PE<br />
Remove need for full mesh of LDP-peers/LSPtunnels<br />
VLLs over multiple tunnels (of different types)<br />
Simplifying VLL provisioning<br />
PE<br />
CE<br />
145 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-segment Pseudo-wire – How can you use them ?<br />
<strong>Ethernet</strong> VLL with MS-PW<br />
T-PE<br />
T-LDP<br />
MS-PW<br />
T-LDP<br />
T-PE<br />
CE<br />
CE<br />
<strong>MPLS</strong><br />
S-PE<br />
T-LDP<br />
S-PE<br />
<strong>MPLS</strong><br />
CE<br />
T-PE<br />
T-LDP<br />
<strong>MPLS</strong><br />
T-LDP<br />
T-LDP<br />
T-PE<br />
CE<br />
T-LDP<br />
<strong>Ethernet</strong> VLL redundancy across multiple areas<br />
e.g. FRR only available within an area/level<br />
S-PE<br />
T-LDP<br />
<strong>MPLS</strong><br />
T-PE<br />
CE<br />
Inter-domain connectivity<br />
[<strong>Metro</strong> w/RSVP] to [core w/LDP] to [metro w/RSVP]<br />
One device needs PWs to many remote devices<br />
<strong>MPLS</strong> tunnel<br />
146 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-segment Pseudo-wire – How do they work ?<br />
same<br />
TUN-1 PW-1 Customer frame TUN-2 PW-1 Customer frame<br />
Customer frame<br />
PE<br />
P<br />
PE<br />
Single Segment PW<br />
Access<br />
Node<br />
VLL<br />
Access<br />
Node<br />
swapped<br />
TUN-1 PW-1 Customer frame TUN-2 PW-2 Customer frame<br />
Customer frame<br />
T-PE<br />
S-PE<br />
T-PE<br />
Multi Segment PW<br />
Access<br />
Node<br />
VLL<br />
Access<br />
Node<br />
147 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Multi-segment Pseudo-wire – Redundancy<br />
Inter-metro/domain Redundant <strong>Ethernet</strong> VLLs with MS-PW<br />
–Individual segments can have <strong>MPLS</strong> (FRR…) protection<br />
–Configure parallel MS-PW for end-end protection<br />
Domain A<br />
Inter-domain<br />
Domain B<br />
S-PE<br />
S-PE<br />
<strong>MPLS</strong><br />
<strong>MPLS</strong><br />
<strong>MPLS</strong><br />
CE<br />
T-PE<br />
S-PE<br />
S-PE<br />
Active Active Active<br />
T-PE<br />
CE<br />
Endpoint with 2 PWs with<br />
preference determining TX<br />
Endpoint with 2 PWs with<br />
preference determining TX<br />
148 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
5. Summary<br />
149 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Summary<br />
<br />
<br />
<br />
<br />
<strong>Ethernet</strong> Services are in a period of tremendous growth with great<br />
revenue potential for service providers<br />
The <strong>Metro</strong> <strong>Ethernet</strong> Forum has standardised <strong>Ethernet</strong> services and<br />
continues to enhance specifications<br />
Traditional forms of <strong>Ethernet</strong> delivery are no longer suitable for the<br />
delivery of “carrier-grade” <strong>Ethernet</strong> services<br />
<strong>MPLS</strong> provides a proven platform for the delivery of scalable, flexible,<br />
feature-rich <strong>Ethernet</strong> services using the same infrastructure used to<br />
deliver other <strong>MPLS</strong>-<strong>based</strong> services<br />
150 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
6. Questions ???<br />
151 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
Thank You<br />
152 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>
www.alcatel-lucent.com<br />
153 | <strong>MPLS</strong>-<strong>based</strong> <strong>Metro</strong> <strong>Ethernet</strong> <strong>Networks</strong>, February <strong>2011</strong>